Thanks to the advances in the Internet of Things (IoT), Condition-based Maintenance (CBM) has progressively become one of the most renowned strategies to mitigate the risk arising from failures. Within any CBM framework, non-linear correlation among data and variability of condition monitoring data sources are among the main reasons that lead to a complex estimation of Reliability Indicators (RIs). Indeed, most classic approaches fail to fully consider these aspects. This work presents a novel methodology that employs Accelerated Life Testing (ALT) as multiple sources of data to define the impact of relevant PVs on RIs, and subsequently, plan maintenance actions through an online reliability estimation. For this purpose, a Generalized Linear Model (GLM) is exploited to model the relationship between PVs and an RI, while a Hierarchical Bayesian Regression (HBR) is implemented to estimate the parameters of the GLM. The HBR can deal with the aforementioned uncertainties, allowing to get a better explanation of the correlation of PVs. We considered a numerical example that exploits five distinct operating conditions for ALT as a case study. The developed methodology provides asset managers a solid tool to estimate online reliability and plan maintenance actions as soon as a given condition is reached.

Over the last few decades, reliability analysis has attracted significant interest due to its importance in risk and asset integrity management. Meanwhile, Bayesian inference has proven its advantages over other statistical tools, such as maximum likelihood estimation (MLE) and least square estimation (LSE), in estimating the parameters characterizing failure modelling. Indeed, Bayesian inference can incorporate prior beliefs and information into the analysis, which could partially overcome the lack of data. Accordingly, this paper aims to provide a closed-mathematical representation of Bayesian analysis for reliability assessment of industrial components while investigating the effect of the prior choice on future failures predictions. To this end, hierarchical Bayesian modelling (HBM) was tested on three samples with distinct sizes, while five different prior distributions were considered. Moreover, a beta-binomial distribution was adopted to represent the failure behavior of the considered device. The results show that choosing strong informative priors leads to distinct predictions, even if a larger sample size is considered. The outcome of this research could help maintenance engineers and asset managers in integrating their prior beliefs into the reliability estimation process.

Safety Improvement of engineering processes, especially Oil & Gas operations, has gained a lot of attention during the last decades. This fundamental vision results in risk remediation programs, minimizing the risks of failure, and reducing the associated costs for operation and maintenance. As failures may represent serious threats for both humans and the environment, a comprehensive tool is required to employ maintenance and avoid immoderate dangerous consequences. Traditional risk frameworks mainly include estimation approaches, such as Fault Tree (FT) and Event Tree (ET), producing more simplified models than other tools, such as Bayesian inference. The present work aimed at developing an advanced Risk-Based Maintenance (RBM) methodology for prioritizing maintenance operations, by addressing associated uncertainties through the accident modelling of the process. For this purpose, a Hierarchical Bayesian Approach (HBA) is applied to estimate the failure probabilities of each component while a Failure Mode, Effects and Criticality Analysis is performed to assess the severity. With Markov Chain Monte Carlo simulation from likelihood function and prior distribution, the HBA is capable of incorporating the fluctuations and uncertainties associated with operational data including the variability between the source of data and the correlation of observations. Lastly, to make a meaningful difference between different kinds of risk consequences, whether the risk has a direct or indirect loss, the cost of failures of components is accounted for. To demonstrate the application of the methodology, a Natural Gas Reduction and Measuring Station (NGRMS) is taken into account as a case study. The outcome of the case study proofed that PTG and pump are the most failures sensitive components among other if they being left unattended in the operation with an average number of failure occurrences of 67 and 45; While the THT pipelines and THT tank are less sensitive for being considered for major maintenance request with almost average of 5 times in their lifetime. The proposed method can be exploited by maintenance engineers, asset managers, and policymakers to reduce the downtime periods as well as the risks of on-going operations.

During the last decades, the optimization of the maintenance plan in process plants has lured the attention of many researchers due to its vital role in assuring the safety of operations. Within the process of scheduling maintenance activities, one of the most significant challenges is estimating the reliability of the involved systems, especially in case of data scarcity. Overestimating the average time between two consecutive failures of an individual component could compromise safety, while an underestimate leads to an increase of operational costs. Thus, a reliable tool able to determine the parameters of failure modelling with high accuracy when few data are available would be welcome. For this purpose, this paper aims at comparing the implementation of three practical estimation frameworks in case of sparse data to point out the most efficient approach. Hierarchical Bayesian modelling (HBM), maximum likelihood estimation (MLE), and least square estimation (LSE) are applied on data generated by a simulated stochastic process of a natural gas regulating and metering station (NGRMS), which was adopted as a case of study. The results identify the Bayesian methodology as the most accurate for predicting the failure rate of the considered devices, especially for the equipment characterized by less data available. The outcomes of this research will assist maintenance engineers and asset managers in choosing the optimal approach to conduct reliability analysis either when sufficient data or limited data are observed.